Revert "Fix: Dockerfile"

This reverts commit 5dcd1567af5849e91b8f052e85eab37f8b44b03e.

Dockerfile

@@ -1,12 +1,11 @@
-FROM pytorch/pytorch:2.3.0-cuda12.1-cudnn8-runtime
+FROM python:3.10-slim
 
 WORKDIR /app
 
 # Install curl for healthcheck
 RUN apt-get update && apt-get install -y curl && rm -rf /var/lib/apt/lists/*
 
-# torch + CUDA 12.1 + cuDNN 8 are already in the base image.
-# requirements.txt installs only the remaining app-level deps.
+# Install dependencies first (layer cache optimization)
 COPY requirements.txt .
 RUN pip install --no-cache-dir -r requirements.txt
 

HANDOVER.md

@@ -0,0 +1,211 @@
+# AgentDebuggerEnv — Project Handover
+
+## What This Project Is
+A GRPO-trained LLM (Qwen2.5-Coder-7B-Instruct) that learns to debug Python code through
+structured hypothesis-driven reasoning. Submitted to the Meta + PyTorch + HuggingFace OpenEnv Hackathon.
+
+---
+
+## Repo & Remotes
+
+| Remote | URL |
+|---|---|
+| GitHub (source of truth) | https://github.com/shasshaank/meta_hackthon.git |
+| HF Training Space | https://huggingface.co/spaces/shashaank0707/AgentDebugger-training-v2 |
+| HF Trained Model | https://huggingface.co/shashaank0707/AgentDebugger-trained |
+
+Push to GitHub first, then to HF Space if needed:
+```bash
+git push origin main
+git push space main --force   # space remote = HF training space
+```
+
+The `space` remote URL includes your HF token:
+```
+https://shashaank0707:YOUR_HF_TOKEN@huggingface.co/spaces/shashaank0707/AgentDebugger-training-v2
+```
+
+---
+
+## Project Structure
+
+```
+meta_hackathon/
+├── app.py                      # Gradio training monitor — launched by HF Space SDK
+├── training/
+│   └── train_grpo.py           # Main training script (GRPO via TRL)
+├── server/
+│   ├── reward_calculator.py    # Multi-component reward (format, hypothesis, fix, semantic)
+│   ├── models.py               # parse_agent_output() — parses structured LLM output
+│   └── app.py                  # FastAPI server (for the inference/env Space, not training)
+├── data/
+│   ├── bugs_tier1.jsonl        # 9 easy bugs (used steps 0–150)
+│   ├── bugs_tier2.jsonl        # 31 medium bugs (added at step 150)
+│   ├── bugs_tier3.jsonl        # 21 hard bugs (added at step 350 → was 600)
+│   └── generate_bugs.py        # Script that generated the bug datasets
+├── requirements.txt            # HF Space deps (gradio[oauth,mcp]==6.13.0, cu121 torch)
+├── requirements_kaggle.txt     # Kaggle/RunPod deps (no torch pin, bitsandbytes==0.45.3)
+├── inference.py                # Inference wrapper for evaluation
+├── Dockerfile                  # For the inference/env Space (not the training space)
+└── README.md                   # HF Space config header (sdk: gradio, app_file: app.py)
+```
+
+---
+
+## Dependency Versions (locked — do not change without testing)
+
+| Package | Version | Why pinned |
+|---|---|---|
+| `trl` | `0.14.0` | First version with `GRPOTrainer` + `GRPOConfig` |
+| `pydantic` | `2.12.5` | Only version satisfying both gradio base AND gradio[mcp] constraints |
+| `gradio` | `6.13.0[oauth,mcp]` | HF Space builder requires extras in one install pass |
+| `bitsandbytes` | `0.45.3` (Kaggle) / `0.43.3` (HF Space cu121) | 0.45.3 has CUDA 12.x binaries; 0.43.3 works with cu121 |
+| `transformers` | `4.46.3` | Tested with TRL 0.14.0 |
+| `torch` | `2.5.1+cu121` (HF Space) / pre-installed (Kaggle) | |
+
+**GRPOConfig param name:** `max_completion_length` (NOT `max_new_tokens` — that's the old name, breaks on 0.14.0)
+
+---
+
+## Training Script — Key Design Decisions
+
+### GPU Auto-Detection (train_grpo.py ~line 260)
+The script detects GPU at runtime and sets all hyperparams automatically:
+
+| GPU | dtype | batch | grad_accum | num_gen | max_comp | lora_r |
+|---|---|---|---|---|---|---|
+| A100 40GB+ | bfloat16 | 2 | 4 | 8 | 256 | 16 |
+| V100 32GB | float16 | 1 | 8 | 6 | 220 | 12 |
+| T4 / ≤16GB | float16 | 1 | 8 | 4 | 160 | 8 |
+
+**Critical:** P100 is NOT supported — PyTorch 2.x dropped sm_60 support. Use T4 instead.
+
+### Curriculum
+- Steps 0–150: Tier 1 bugs only (9 bugs)
+- Steps 150–350: Tier 1 + Tier 2 (40 bugs)
+- Steps 350+: All tiers (61 bugs)
+
+### Reward Components (server/reward_calculator.py)
+| Component | Weight | What it measures |
+|---|---|---|
+| format_compliance | 0.10 | All 5 fields present (OBSERVATION/HYPOTHESIS/CONFIDENCE/ACTION/DETAIL) |
+| hypothesis_quality | 0.20 | Length + references specific variable names |
+| localization | 0.15 | Correct function/line identified |
+| fix_quality | 0.35 | Tests pass on proposed fix |
+| semantic_similarity | 0.10 | Similarity to canonical fix |
+| efficiency_potential | 0.10 | Potential-based shaping (Ibrahim et al. 2024) |
+
+### Required Output Format
+```
+OBSERVATION: [specific observations with line numbers]
+HYPOTHESIS: [2+ sentences explaining root cause with variable names]
+CONFIDENCE: [low | medium | high]
+ACTION: [inspect_lines | run_tests | propose_fix | request_context | give_up]
+DETAIL: [complete fixed function code if propose_fix, else details]
+```
+
+---
+
+## Running Training
+
+### On Kaggle (T4 — free):
+```python
+# Cell 1 — install
+!pip install -q wandb==0.18.7 datasets==3.0.2 transformers==4.46.3 \
+    accelerate==1.0.1 trl==0.14.0 bitsandbytes==0.45.3 peft==0.13.2
+
+# Cell 2 — clone + secrets
+from kaggle_secrets import UserSecretsClient
+import os
+secrets = UserSecretsClient()
+os.environ["WANDB_API_KEY"] = secrets.get_secret("WANDB_API_KEY")
+os.environ["HF_TOKEN"]      = secrets.get_secret("HF_TOKEN")
+!git clone https://github.com/shasshaank/meta_hackthon.git /kaggle/working/repo
+%cd /kaggle/working/repo
+
+# Cell 3 — train (streams output live)
+import subprocess, sys
+proc = subprocess.Popen(
+    [sys.executable, "training/train_grpo.py"],
+    stdout=subprocess.PIPE, stderr=subprocess.STDOUT,
+    text=True, bufsize=1, cwd="/kaggle/working/repo"
+)
+for line in proc.stdout:
+    print(line, end="", flush=True)
+proc.wait()
+
+# Cell 4 — save outputs after training
+import shutil
+shutil.copytree("/kaggle/working/repo/checkpoints", "/kaggle/working/checkpoints", dirs_exist_ok=True)
+```
+
+**Kaggle secrets needed:** `WANDB_API_KEY`, `HF_TOKEN`
+**Kaggle GPU:** T4 x1 (NOT P100 — incompatible with modern PyTorch)
+**Expected time:** ~8–10 hours for 500 steps (default max_steps=500)
+
+### On RunPod (A100 — ~$1.09/hr):
+```bash
+git clone https://github.com/shasshaank/meta_hackthon.git && cd meta_hackthon
+pip install -q wandb==0.18.7 datasets==3.0.2 transformers==4.46.3 \
+    accelerate==1.0.1 trl==0.14.0 bitsandbytes==0.45.3 peft==0.13.2
+WANDB_API_KEY=xxx HF_TOKEN=xxx python training/train_grpo.py
+```
+**Expected time:** ~3–4 hours for 1000 steps on A100 40GB
+
+### Resume from checkpoint:
+```bash
+python training/train_grpo.py --resume ./checkpoints/checkpoint-400
+```
+
+### Local sanity check (no GPU):
+```bash
+python training/train_grpo.py --test-local
+```
+
+---
+
+## HF Space Setup (training monitor)
+
+The training Space (`AgentDebugger-training-v2`) is a Gradio app that:
+1. On startup, spawns `training/train_grpo.py` in a background thread
+2. Shows a live training log in the UI, auto-refreshing every 30s
+
+**Required Space secrets:**
+- `WANDB_API_KEY`
+- `HF_TOKEN`
+
+**Push to Space:**
+```bash
+git remote set-url space https://shashaank0707:YOUR_HF_TOKEN@huggingface.co/spaces/shashaank0707/AgentDebugger-training-v2
+git push space main --force
+```
+
+---
+
+## Known Issues Fixed (do not revert)
+
+| Issue | Fix |
+|---|---|
+| `ImportError: cannot import name 'GRPOTrainer'` | `trl==0.12.2` → `trl==0.14.0` |
+| `TypeError: GRPOConfig got unexpected keyword 'max_new_tokens'` | renamed to `max_completion_length` |
+| `pydantic` conflict with `gradio[mcp]` | `pydantic==2.10.6` → `2.12.5` |
+| `P100 not supported by PyTorch 2.x` | Switch to T4 on Kaggle |
+| `bitsandbytes CUDA binary not found` | `bitsandbytes==0.43.3` → `0.45.3` on Kaggle |
+| `unsloth` CUDA driver crash on HF A100 | Replaced with `bitsandbytes + peft` |
+| `gradio every=` deprecation | Replaced with `gr.Timer(value=30)` |
+
+---
+
+## W&B Dashboard
+https://wandb.ai/shashaankjain07-keshav-memorial-college-of-law/AgentDebuggerEnv
+
+Training runs appear here automatically when `WANDB_API_KEY` is set.
+
+---
+
+## What's Left To Do
+
+- [ ] **Finish training** — 500–1000 steps, model pushes to HF Hub automatically on completion
+- [ ] **Verify trained model** — run `inference.py` against the trained model checkpoint
+- [ ] **Update HF Space README** — change curriculum description to match actual step boundaries (150/350)
+- [ ] **Submission** — ensure the inference/env Space (`AgentDebugger-env`) is live and healthy for judging

get_logs.py

@@ -1,11 +0,0 @@
-import urllib.request
-import json
-import sys
-
-url = "https://huggingface.co/api/spaces/shashaank0707/AgentDebugger-training-v3"
-req = urllib.request.Request(url)
-with urllib.request.urlopen(req) as response:
-    data = json.loads(response.read().decode())
-    
-# Not sure where the build logs are in the API, but I can check the state
-print(data.get('runtime', {}).get('stage'))

req_test.txt

@@ -1,9 +0,0 @@
-gradio[oauth,mcp]==6.13.0
-pydantic==2.12.5
-wandb==0.18.7
-datasets==3.0.2
-transformers==4.46.3
-accelerate==1.0.1
-trl==0.14.0
-mergekit
-peft==0.13.2

req_test2.txt

@@ -1,9 +0,0 @@
-gradio[oauth,mcp]==6.13.0
-pydantic
-wandb
-datasets
-transformers
-accelerate
-trl==0.14.0
-mergekit
-peft

req_test3.txt

@@ -1,10 +0,0 @@
-click>=8.1.0
-gradio[oauth,mcp]==6.13.0
-pydantic
-wandb
-datasets
-transformers
-accelerate
-trl==0.14.0
-mergekit
-peft

requirements.txt

@@ -1,4 +1,3 @@
-# torch, CUDA 12.1, and cuDNN 8 are pre-installed in the base image:
-# pytorch/pytorch:2.3.0-cuda12.1-cudnn8-runtime
-# Do NOT add torch here — pip would resolve to the CPU wheel from default PyPI
-# and overwrite the CUDA-enabled torch from the base image.
+# torch must be installed at build time (CUDA wheel is ~2GB, too slow at runtime)
+# Everything else is installed at runtime in training/train_grpo.py
+torch

scratch/normalize_inputs.py

@@ -1,52 +0,0 @@
-import os
-import sys
-import pprint
-
-sys.path.append(os.path.abspath('.'))
-from data.generate_bugs import TIER1_BUGS, TIER2_BUGS, TIER3_BUGS
-
-def normalize_test_cases(bugs):
-    for b in bugs:
-        for t in b.get("test_cases", []):
-            inp = t["input"]
-            if isinstance(inp, (list, tuple)):
-                t["input"] = list(inp)
-            else:
-                t["input"] = [inp]
-
-normalize_test_cases(TIER1_BUGS)
-normalize_test_cases(TIER2_BUGS)
-normalize_test_cases(TIER3_BUGS)
-
-def dump_var(f, name, val):
-    f.write(f'{name} = ')
-    f.write(pprint.pformat(val, sort_dicts=False, width=120))
-    f.write('\n\n')
-
-with open("data/generate_bugs.py", "w", encoding="utf-8") as f:
-    f.write('"""\nAgentDebuggerEnv - Bug Dataset Generator\n\n')
-    f.write('Generates three tiers of buggy Python functions for curriculum learning:\n')
-    f.write('  Tier 1 (easy):   Off-by-one errors, wrong operators, simple logic inversions\n')
-    f.write('  Tier 2 (medium): Incorrect algorithm logic, wrong variable references, subtle type errors\n')
-    f.write('  Tier 3 (hard):   Multi-bug interactions, concurrency, edge-case-only failures\n\n')
-    f.write('Usage:\n  python data/generate_bugs.py\n\n')
-    f.write('Outputs:\n  data/bugs_tier1.jsonl  (~40 bugs)\n  data/bugs_tier2.jsonl  (~30 bugs)\n  data/bugs_tier3.jsonl  (~20 bugs)\n"""\n\n')
-    f.write('import json\nimport os\n\n')
-    
-    dump_var(f, 'TIER1_BUGS', TIER1_BUGS)
-    dump_var(f, 'TIER2_BUGS', TIER2_BUGS)
-    dump_var(f, 'TIER3_BUGS', TIER3_BUGS)
-    
-    f.write('def write_jsonl(bugs: list, path: str):\n')
-    f.write('    with open(path, "w") as f:\n')
-    f.write('        for bug in bugs:\n')
-    f.write('            f.write(json.dumps(bug) + "\\n")\n\n')
-    f.write('if __name__ == "__main__":\n')
-    f.write('    os.makedirs("data", exist_ok=True)\n')
-    f.write('    write_jsonl(TIER1_BUGS, "data/bugs_tier1.jsonl")\n')
-    f.write('    write_jsonl(TIER2_BUGS, "data/bugs_tier2.jsonl")\n')
-    f.write('    write_jsonl(TIER3_BUGS, "data/bugs_tier3.jsonl")\n')
-    f.write('    print(f"Tier 1: {len(TIER1_BUGS)}, Tier 2: {len(TIER2_BUGS)}, Tier 3: {len(TIER3_BUGS)}")\n')
-    f.write('    print("\\nDone. Run training/train_grpo.py to start training.")\n')
-
-print("Normalization applied successfully.")

test_pip.sh

@@ -1,6 +0,0 @@
-#!/bin/bash
-python3 -m venv venv_test
-source venv_test/bin/activate
-pip install pip -U
-pip install datasets "huggingface-hub>=0.30" "hf-transfer>=0.1.4" "protobuf<4" "click<8.1"
-pip install -r requirements.txt gradio[oauth,mcp]==6.13.0 "uvicorn>=0.14.0" "websockets>=10.4" spaces

test_pip2.sh

@@ -1,6 +0,0 @@
-#!/bin/bash
-python3 -m venv venv_test2
-source venv_test2/bin/activate
-pip install pip -U
-pip install datasets "huggingface-hub>=0.30" "hf-transfer>=0.1.4" "protobuf<4" "click<8.1"
-pip install -r req_test.txt gradio[oauth,mcp]==6.13.0 "uvicorn>=0.14.0" "websockets>=10.4" spaces

test_pip3.sh

@@ -1,6 +0,0 @@
-#!/bin/bash
-python3 -m venv venv_test3
-source venv_test3/bin/activate
-pip install pip -U
-pip install datasets "huggingface-hub>=0.30" "hf-transfer>=0.1.4" "protobuf<4" "click<8.1"
-pip install -r req_test2.txt spaces

test_pip4.sh

@@ -1,6 +0,0 @@
-#!/bin/bash
-python3 -m venv venv_test4
-source venv_test4/bin/activate
-pip install pip -U
-pip install datasets "huggingface-hub>=0.30" "hf-transfer>=0.1.4" "protobuf<4" "click<8.1"
-pip install -r req_test3.txt spaces